high speed networks
TRANSCRIPT
HIGH SPEED NETWORKSINTEGRATED SERVICES
&QUEUING DISCIPLINE
Submitted ByMr.V.S.Patil
CONTENTS
• What is ISA• Need of ISA• Internet Traffic• - Elastic Traffic • - Inelastic Traffic• ISA Approach• ISA Functions• ISA Components• ISA Services• Queuing Discipline• -FIFO Queuing• -FQ (Fair Queuing)• -PS (Processor sharing)• -BRFQ (Bit-Round Fair Queuing)• -GPS (Generalized Processor Sharing)• -WFQ (Weighted Fair Queuing)
WHAT IS ISA • Internet IP were designed to provide a best effort fair delivery
services.• Internet treats all packets equally.• As the level of traffic on the network grows and congestion
occurs,all packets delivery is slowed down.• In congestion packets are dropped.• Due to increase in volume and introduction of Real-Time
application, IP are inadequate.• To cope with this it is not enough to increase Internet capacity .• Effective methods for managing the traffic and controlling
congestion are needed.• We have two Traffic Management frame works • - INTEGRATED SERVICES• -DIFFERENTIATED SERVICES.
NEED OF ISA• IPv4 header is equipped with fields that can specify precedence and
type of service.• This information is ignored by ROUTER.• REAL-TIME applications are not well supported by such a
configuration.• Alternative is to use ATM .• But changing from IP to ATM is very costly .• Strong need is to use QoS (Quality of Service) requirement with the
help of TCP/IP architecture.• Fundamental requirement is to add new functionality to ROUTER.• To meet this IETF is developing a suit of standard under the umbrella
of ISA (Integrated Services Architecture).• ISA intended to provide QoS transport over IP – based internet.
INTERNET TRAFFIC
• Traffic on internet can be divided into two broad catagories:
- ELASTIC TRAFFIC - INELASTIC TRAFFIC.
ELASTIC TRAFFIC • Elastic traffic can adjust over wide range, to changes in delay and
throughput.
• This is traditional type of traffic supported on IP-based internet.
• Applications that can be classified as elastic include common application that operate on TCP or UDP , includes
- E-mail : is generally insensitive to changes in delay.
- FTP : when file transfer is done on-line ,user expects the delay to be proportional to the file size.
- Web access are quite sensitive to delay.
INELASTIC TRAFFIC• Inelatic traffic does not easily adapt to changes in delay and throughput across
the internet.
• Example is REAL-TIME application.
• Requirements for INELASTIC TRAFFIC may include the following
- THROUGHPUT: A minimum Throughput value is required .
- DELAY : An ex. Of delay sensitive application is stock Market, someone who receives later service will act later ,that is a disadvantage.
- JITTER : The larger the allowable delay variation, the longer the real delay in delivering the data.Upper bound on Jitter is required.
- PACKET LOSS :Real time applications vary in amount of packet loss ,if any ,they can sustain.
• This requirements are difficult to meet in an environment with variable queuing delays and congestion losses.
• the internet architecture Inelastic traffic introduces two new requirements
1) Some means is needed to give preferential treatment to application with more demanding requirements.
2) Elastic traffic must still be supported ,so in times of congestion, inelatic traffic will continue to supply a high load and elastic traffic will be crowded off the internet.
ISA Approach• The purpose of ISA is to enable the provision of QoS support over
IP-based internets .• The central design issue for ISA is how to share the available
capacity in times of congestion.• In IP-based internet ,tools for controlling congestion and providing
services are limited.
• Routers have two mechanisms to work with: - ROUTING ALGORITHM: Most routing protocols in use in internet allow routes to be
selected to minimize delay. - PACKET DISCARD : When a router’s buffer over flows ,it
discards packets ,the most recent packet is discarded.
ISA Functions
ISA makes use of following functions to manage congestion and provide QoS transport :
• Admission Control• Routing Algorithm• Queuing discipline• Discard Policy.
ISA COMPONENTS
RoutingDatabase
QoS Queuing
Best-Effort Queuing
RoutingProtocols
ReservationProtocol Admission
ControlManagement
Agent
TrafficControlDatabase
Classifier &RouteSelection
PacketScheduler
BACKGROUND FUNCTIONS
FORWARDING FUNCTIONS
ISA Components• RESERVATION PROTOCOL:
This protocol is used among Routers and between routers and end systems to reserve resources for a new flow. RSVP protocol is used for this purpose.
• ADMISSION CONTROL :
When a new flow is requested ,the RESERVATION PROTOCOL invokes the ADMISSION CONTROL function ,it determines if sufficient resources are available for the flow.
• MANAGEMENT AGENT :
It is able to modify the traffic control database and to direct admission control policies.
• ROUTING PROTOCOL : It is responsible for maintaining a routing database.
• CLASSIFIER & ROUTE SELECTION:
• For the purpose of forwarding and traffic control ,incoming packets must be mapped into classes.
• PACKET SCHEDULER :
It manages one or more queues for each output port.It determines the order in which packets are transmitted.
ISA Services Currently three catagories of services are
defined:• Guaranteed• Controlled load• Best effort
Queuing Discipline• The important component of an ISA implementation is the Queuing
discipline used at the routers.• Routers traditionally have used a FIFO queuing discipline also
known as FCFS.• When a new packet arrives it is placed at the end of the queue.
Xmit
Flow 1
Flow 2
Flow N
Multiplexed
Output
FIFO
Drawbacks Of FIFO
• No special treatment is given to packets from flows that are of higher priority or more delay sensitive.
• If smaller packets are queued behind a long packet ,then flows of larger packets get better service.
FAIR QUEUING (FQ)• Router maintains multiple queues at each output port.
• Queues are serviced in Round Robin fashion, taking one packet from each non empty queue in turn.
Xmit
Flow 1
Flow 2
Flow N
Multiplexed Output
PROCESSOR SHARING (PS)• In FQ short packets are penalized.• More capacity goes to flows with longer packet size.• This is overcome with BIT-ROUND FAIR QUEUING (BRFQ).• It transmit only one bit from each queue on each round.• If there are N queues then each queue receives exactly 1/N of the available capacity.• This BIT-by-BIT approach is called as PS. R(t) = No. of rounds occurred in time t.
N(t) = No. of non empty queues at time t. Pi = Transmission time for packet i in queue . Ti = Arrival time for packet i in queue. Si = Value of R(t) when packet i in queue begins transmission.
Fi = Value of R(t) when packet i in queue ends transmission.
BIT – ROUND FAIR QUEUING(BRFQ)
• To transmit entire packet rather than individual bits.
• BRFQ is designed for a bit-by-bit Round-Robin discipline.
• Whenever a packet finishes transmission, the next packet sent is the one with the smallest value of Fi.
Flow 1
Flow 2
Flow 3
12 3 21
2 3 2 1 2 3 2FIFO
FQ
BRFQ
Same as FIFO
2 3 1 2 2 3 1 2
GENERALIZED PROCESSOR SHARING (GPS)
• BRFQ is an improvement over FQ or FIFO, in that it fairly allocates the available capacity among all active flows.
• It is not able to provide different amounts of capacity to different flows.
• To support QoS transport ,this differential allocation capability is needed.
• With GPS each flow is assigned a weight that determines how many bits are transmitted from that queue during each round.
• If the weight for a given flow is 5, then during each round that the queue is non empty, 5 bits will be transmitted.
WEIGHTED FAIR QUEUING (WFQ)
• WFQ emulates the bit-by-bit GPS.
• Whenever a packet finishes transmission, the next packet sent is the one with smallest value of Fi.
• GPS is attractive because it enables a router to assign approprite weight to each flow to guarentee service & because it is possible to guarentee an upper bound on delay.
CONCLUSION• ISA provides effective methods for managing the traffic
and controlling the congestion.
• RSVP and OSPF protocols are used to make reservation and select routes based on QoS.
• Queuing discpline ,a vital element of ISA is an effective queuing policy that takes into account the different requirements of different flows.
• Discard policy determines which packet to transmit next if a number of packets are queued for the same port.
THANK YOU.